To fully grasp the public health importance of cigarette smoking, we must first be able to measure its impact on mortality at the societal level. While the impact of smoking on mortality at the individual level is well- established, examining the population-level impact poses greater challenges. Current methods for determining the mortality burden of smoking in a population are informative, but suffer from several inadequacies that limit our understanding of the contribution of smoking to differences in mortality experiences across societies. This project will expand our knowledge of the importance of smoking for population health and demonstrate its potential to be an important factor explaining population-level trends in mortality. This project has three objectives. First, we investigate the impact of smoking on mortality in a population that exhibits extremely low rates of cigarette smoking - Mexican-Americans. We expect that low cigarette consumption in this population will translate into relatively low smoking-related mortality compared with non-Hispanic whites. Thus, smoking may provide a convincing answer to the "Hispanic Paradox". We will model all cause mortality as a function of smoking behavior and estimate the contribution of cigarette smoking to the mortality advantage of Mexican-Americans. The second objective is to identify the best method for predicting national trends in lung cancer mortality in developed countries. In the absence of detailed cohort smoking histories, researchers are often forced to measure population smoking behavior using single-year period measures. Most studies use the prevalence of cigarette smoking in the population, which is ascertained by self reports in representative surveys. Annual cigarette sales are used less often, but it is unknown which is the superior measure of population smoking behavior. We will compare the performance of each measure in predicting national lung cancer mortality from 1950 through 2005 in twenty developed countries. The third objective is to develop a new method for estimating the number of deaths attributable to smoking in developed countries. Our method improves upon previous methods in that it can be applied to a greater number of contexts and it does not rely heavily on assumptions about the individual-level relationship between smoking and mortality. We will investigate the macro-level relationship between lung cancer mortality and mortality from all other causes of death. This method will improve the accuracy of previous attributable risk estimates and will allow us to calculate attributable risk in a greater number of populations than ever before. PUBLIC HEALTH RELEVANCE: To fully grasp the public health importance of cigarette smoking, we must first be able to measure its impact on mortality at the societal level. This project seeks to improve our understanding of the significance of smoking for the health of populations and elucidate its potential to explain differences in life expectancy across places.